COMMUNICATION SYSTEMS

TECHNICAL FIELD

[001] Disclosed are embodiments related to non-ortho gonal multiple access (NOMA) communication systems.

BACKGROUND

[002] The design of multiple access schemes is of interest in the design of cellular telecommunication systems. The goal of multiple access schemes is to provide multiple user equipments (UEs) (i.e., wireless communication devices, such as, for example, smartphones, tablets, phablets, smart sensors, wireless Intemet-of-Things (IoT) devices, etc., that are capable of wirelessly communicating with an access point) with radio resources in a spectrum, cost, and complexity-efficient manner. In 1G-3G wireless communication systems, frequency division multiple access (FDMA), time division multiple access (TDMA) and frequency division multiple access (CDMA) schemes have been introduced. Long-Term Evolution (LTE) and LTE-Advanced employ orthogonal frequency division multiple access (OFDMA) and single-carrier (SC)-FDMA as orthogonal multiple access (OMA) schemes. Such orthogonal designs have the benefit that there is no mutual interference among UEs, leading to high system performance with simple receivers.

[003] Recently, non-orthogonal multiple access (NOMA) has received considerable attention as a promising multiple access technique for LTE and 5G systems. With NOMA, two or more UEs may share the same radio resources (e.g., time resources, frequency resources, and/or code resources). Particularly, 3GPP has considered NOMA in different applications. For instance, NOMA has been introduced as an extension of the network -assisted interference cancellation and suppression (NAICS) for intercell interference (ICI) mitigation in LTE Release 12 as well as a study item of LTE Release 13, under the name of“Downlink multiuser superposition transmission.” Also, in recent 3GPP meetings, it is decided that new radio (NR) should target to support (at least) uplink NOMA, in addition to the OMA approach. SUMMARY

[004] NOMA exploits channel difference between or among UEs to improve spectrum efficiency. Generally, the highest gain of NOMA is observed in the cases where a“strong” UE (i.e., a UE experiencing a good channel condition with a base station, such as, for example, a UE located in the center of a cell) and a“weak” UE (i.e., a UE having a poor channel condition with the base station, such as, for example, a UE located at or near a cell edge) are grouped (i.e., use the same radio resources). However, the implementation of NOMA implies: 1) use of more advanced and complex receivers to enable multiuser signal separation, 2) more difficult synchronization, and 3) a higher signal decoding delay

[005] For example, considering downlink NOMA, the strong UE typically uses successive interference cancellation (SIC) to first decode and remove the signal of the weak UE and then decode its own signal interference-free. As a result, compared to conventional OMA scheme, NOMA-based data transmission leads to higher receiver complexity. Also, compared to OMA -based systems, the two-step decoding process of the strong UE may lead to larger end-to-end transmission delay for the strong UE, as well as for the weak UE (e.g. in scenarios in which their signals should be synchronized). Also, there is a probability that the strong UE cannot correctly decode the message of the weak UE affecting the successful decoding probability of its own message.

[006] Also, while using NOMA outperforms OMA in terms of sum rate, the sum rate gain of NOMA is at the cost possible rate loss for the weak UE (e.g., the cell-edge UE). This is because, with downlink NOMA, the weak UE considers the signal of the strong UE as interference and uses the typical OMA -based decoder to decode its own message. Thus, depending on the interference power, there is a probability that NOMA-based transmission leads to lower achievable rates for the weak UE, compared to the cases using OMA. Also, there may be unfair resource allocation between the weak and the strong UEs.

[007] In one embodiment, this disclosure describes an adaptive receiver for the weak

UEs using NOMA-based schemes. The objective is to improve the achievable rate of the weak UE and improve the fairness among the grouped UEs. At the same time, the proposed scheme gives the chance to reduce the error probability of the strong UE. Thus, compared to conventional NOMA, the proposed scheme increases the network sum throughput. In the proposed scheme, depending on the message decoding status of the strong UE, different decoding schemes may be considered by the weak UE such that it can remove the interfering signal of the strong UE and, thereby, improve its own achievable rate. Compared to the conventional NOMA techniques, the proposed scheme increases the throughput of the weak UEs and improves the fairness among the grouped UEs. Moreover, the proposed scheme makes it possible to improve the performance of the strong UE and, consequently, increase the network sum throughput.

[008] In another embodiment, this disclosure describes an adaptive receiver for the strong UEs using hybrid automatic repeat request (HARQ) in NOMA -based networks. The objective is to reduce the implementation complexity, the decoding delay, and the probability of error propagation for the strong UEs. At the same time, the proposed scheme reduces the end-to-end transmission delay of the weak UEs as well. In the proposed setup, depending on the message decoding status of the weak UE, different decoding schemes may be considered by the strong UE. Then, depending on the selected decoding scheme, the network node (NN) can synchronize the signals of the UEs correspondingly. Compared to the conventional NOMA techniques, the proposed scheme reduces the receiver complexity of the strong UE

considerably. Also, the proposed scheme reduces the end-to-end transmission delay of the network as well as the error propagation at the strong UE.

[009] Accordingly, in aspect there is provided a method performed by a network node serving a first UE and a second UE. The method includes the network node transmitting, during a first time slot (tl), a first superimposed signal (S(tl)) comprising a first message for the first UE and a second message for the second UE. The method also includes the network node receiving a message transmitted by the first UE indicating that the first UE was unable to decode the first message. The method further includes, after receiving the message transmitted by the first UE indicating that the first UE was unable to decode the first message, the network node providing to the second UE a message indicating that the first UE was unable to decode the first message, thereby enabling the second UE to select an appropriate decoding scheme to use in a subsequent time slot. [0010] In some embodiments, the method further includes, after receiving the message transmitted by the first UE indicating that the first UE was unable to decode the first message, the network node transmitting, during a second time slot (t2), a second superimposed signal (S(t2)) comprising the first message for the first UE and a third message for the second UE.

[0011] In some embodiments, the method further includes, the network node

transmitting, during a third time slot, a third superimposed signal comprising a third message for the first UE and a fourth message for the second UE; the network node receiving a message transmitted by the second UE indicating that the second UE was unable to decode the fourth message for the second UE included in the third superimposed signal; and after receiving the message transmitted by the second UE indicating that the second UE was unable to decode the fourth message, the network node providing to the first UE a message indicating that the network node will retransmit in a fourth time slot at least a portion of the fourth message. In some embodiments, the method also includes, after receiving the message transmitted by the second UE indicating that the second UE was unable to decode the fourth message, the network node transmitting, during the fourth time slot, a fourth superimposed signal comprising the fourth message for the second UE and a fifth message for the first UE.

[0012] In another aspect there is provided a method performed by a first UE that is being served by a network node. The method includes the first UE receiving, during a first time slot, a first superimposed signal, the first superimposed signal comprising a first message for the first UE and a second message for a second UE being served by the network node. The method also includes, after receiving the first superimposed signal, the first UE obtains information that indicates that the second UE was unable to decode the second message and uses the obtained information to decode a third message for the first UE that is contained in a second superimposed signal received by the first UE.

[0013] In some embodiments, obtaining the information comprises the first UE receiving a message transmitted by the network node, wherein the message indicates that the second UE was unable to decode the second message. In other embodiments, obtaining the information comprises the first UE receiving a negative acknowledgement transmitted by the second UE. [0014] In some embodiments, the method further includes: the first UE decoding the first superimposed signal to obtain from the first superimposed signal the first message for the first UE; the first UE using the obtained first message and the first superimposed signal to generate a residual signal comprising the second message for the second UE; and the first UE buffering the residual signal. In some embodiments, the method further includes the first UE receiving, during a second time slot, a second superimposed signal, the second superimposed signal comprising a third message for the first UE and at least a portion of the second message; the first UE using the buffered residual signal and the received second superimposed signal to obtain said at least a portion of the second message. In some embodiment, the method also includes the first UE using the obtained second message and the received second superimposed signal to obtain the third message from the second superimposed signal.

BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings, which are incorporated herein and form part of the specification, illustrate various embodiments.

[0016] FIG. 1 illustrates a network node communicating simultaneously with a first UE and a second UE.

[0017] FIG. 2 illustrates processing that occurs during a time slot.

[0018] FIG. 3 illustrates processing, according to one embodiment, that occurs during first and second time slots.

[0019] FIG. 4 illustrates processing, according to another embodiment, that occurs during first and second time slots.

[0020] FIG. 5 is a flow chart illustrating a process according to one embodiment.

[0021] FIG. 6 is a flow chart illustrating a process according to one embodiment.

[0022] FIG. 7 is a block diagram of a network node according to one embodiment.

[0023] FIG. 8 is a diagram showing functional units of a network node according to one embodiment.

[0024] FIG. 9 is a block diagram of a UE according to one embodiment. [0025] FIG. 10 is a diagram showing functional units of a UE according to one embodiment.

[0026] FIG. 11 is a diagram showing functional units of a UE according to one embodiment.

DETAILED DESCRIPTION

[0027] FIG. 1 illustrates a network 100 having a network node (NN) 105 (e.g., a system comprising a 4G or 5G base station or other access point) serving two UEs: UE 101 and UE 102. The two UEs have different channel (or“link”) qualities. In this scenario, UE 102 is a “weak” UE (e.g., a cell-edge UE) and UE 101 is a“strong” UE (e.g. a cell-center UE).

[0028] With respect to uplink OMA transmissions, the UE 101’s and UE l02’s signals are transmitted in orthogonal resources, for instance at the same time but in different frequency bands, and the NN 105 decodes the two transmitted signals separately. With respect to downlink OMA transmissions, the NN 105 transmits for UE 101 a first signal using for example a first frequency band and transmits for UE 102 a second signal using for example a second frequency band that does not overlap with the first frequency band.

[0029] With respect to uplink NOMA, on the other hand, the UEs share the same frequency (or“spectrum”), time resources, and code or spreading resources, if any, to send their messages simultaneously. In such a NOMA scenario, the NN, using for example a SIC receiver, first decodes the message of UE 101 (the“strong” UE), considering the signal of UE 102 as noise. Then, after successfully decoding UE 101’s signal, the NN subtracts UE 101’s signal from the received signal and decodes UE l02’s signal with no interference from UE 101.

[0030] Likewise, with respect to downlink NOMA, UE 101 and UE 102 are served by the NN in common spectrum resource blocks, i.e., time -frequency chunks. We shall consider a frequency slot so that the time-frequency chunks refer to different time slots. Then, with no loss of generality, suppose that UE 101 experiences a better channel quality compared to UE 102 (i.e., UE 101 is the strong UE and UE 102 is the weak UE). That is, we have |h ₂ | < |hi|, where hi represents the channel coefficient of the NN-UE 101 link and h ₂ represents the channel coefficient of the NN-UE 102 link. We define the channel gains as gi = |hi| ² , i= 1 ,2. [0031] Using NOMA, in time slot t the NN generates and transmits a superimposed signal 5(t) = + ^/ ¾A ₂ (t) to both UEs in the same resources. Here, X ₁ (t) and X ₂ (J) are the unit-variance messages for UE 101 and UE 102, respectively, and P*, i = 1, 2, are their corresponding transmit powers with P ₁ + P ₂ = P where P is the NN total power. In this way, the signal received by UE 101 (i.e., Y i(t)) and the signal received by UE 102 (i.e., Y2(t)) is given by:

Yi(t = hiU x i) + Vi¾ ( ) + zm, i = 1, 2, (1)

where

Zi(t) denotes a noise signal (e.g., Gaussian white noise).

[0032] In the above scenario, which is illustrated in FIG. 2, UE 101 uses a SIC receiver to first decode-and-remove the message of UE 102 (i.e., X ₂ ) and then obtain its own message (Xi) with no interference. The UE with the worse channel quality, i.e., UE 102 uses typical decoders to decode its own message in the presence of interference of the signal for UE 101.

[0033] The goal of each UE is to decode its own message, although they may decode the message of the other UE to reduce the interference. With conventional NOMA, UE 102 considers the signal for UE 101 as interference and uses OMA -based receivers to decode its own message. This is because it can be theoretically shown that there is no chance that UE 102 can first decode-and-remove the message of UE 101 (and then, decode its own message interference-free). UE 101, on the other hand, uses a SIC receiver to first decode-and-remove the message of UE 102 and then decode its own message interference-free.

[0034] Compared to conventional OMA -based receivers, SIC is a high-complexity scheme. Also, because the desired signal is decoded in two steps, SIC implies larger decoding delay which affects, e.g., the HARQ feedback process and, thereby, may increase the end-to- end transmission delay for both UEs in the situations where UE l02’s signal should be synchronized with the signal of the UE 101 (different methods can be applied to synchronize the signals - for instance, some sleeping period may be considered by UE 102 (as illustrated in FIG. 2) or the NN may synchronize the signals of the UEs). Finally, with SIC, there is a probability of error propagation. This is because, if the message of UE 102 is not correctly decoded in the first step, the interference is not removed which reduces the probability that the cell-center can successfully decode its own message.

[0035] With this setup, the achievable rate for UE 101 (i.e., Ri) and the achievable rate for UE 102 (i.e., R ₂ ) is given by: + Piflfi),

(2)

i + Pig ₂ '

[0036] From (2), it can be seen that, depending on the received interference power term

P ₁ g ₂ , UE 102 may suffer from low transmission rates. Also, note that, with the message of UE 102 added as an interference, UE 101 suffers from error propagation effect. Particularly, there is a probability that UE 101 can not successfully remove the interference signal affecting the probability that it can decode its own message correctly. Thus, there is reasonably high probability that UE 101 can not decode its own message correctly, and it may need

retransmissions.

[0037] Accordingly, it would be useful to develop a first adaptive decoding scheme where hybrid automatic repeat request (HARQ)-based signals are used to decode the interference seen by UE 102 before UE 102 decodes its message of interest (in this way, removing the interference allows the NN to serve UE 102 with higher rates), and it would also be useful to develop an adaptive decoding scheme where SIC is used only when it is required.

[0038] In the following descriptions, UE 101 is referred to as UEi and UE 102 is referred to as UE ₂ .

[0039] A. The Adaptive Decoding Scheme for the weak UE (UE2)

[0040] Consider two successive time slots t and t (also referred to as tl and t2), as illustrated in FIG. 3. In the proposed setup, UE ₁ uses the same SIC -based decoding scheme as in conventional NOMA. However, depending on the UEi’s message decoding status, different decoding schemes may be considered by UE ₂ in subsequent time slots (e.g. slot t). For simplicity of presentation, assume that, in Slot t, UE ₁ has not decoded its own message Xft) correctly, independently of whether or not it can decode X ₂ (f). Also assume that UE ₂ has correctly decoded its own message X ₂ (f) but, of course, it can not decode Xft) correctly. However, the method presented below is applicable for other message decoding conditions at UE- _L and UE ₂ as long as UE ₁ can not decode X ₁ (t) and needs a retransmission.

[0041] In slot t, UE ₂ first decodes its own message in the presence of interference and then, removes its own message, i.e., E _Zl [R _z ^~ C _z (t) in (1), from the received signal Y _z (t), thereby producing a residual received signal, i.e., T ₂ ( ⁼ b-2-^ ^~ P A( (t) (note that, given that

UE- _L has not correctly decoded X ₁ (t) and with \h ₁ | > \h ₂ \, UE ₂ has no chance to correctly decode X (t) in slot t). UE ₂ then buffers the residual signal. The residual signal gives UE _Z the chance to decode-and-remove the interference of UE-^ in the next time slot (i.e., slot t).

[0042] For instance, in slot t when the NN retransmits message Xi(t) (i.e., the message for UEi that UEi was not able to decode in time slot t), UE _Z uses the buffer residual signal (i.e., Y _z (t) and the signal received in time slot t (i.e., Y _z (t) = first combine the two copies of the UE- _L signal and decode-and-remove it from Y ₂ (t). Then, UE _Z decodes X _z (t) interference-free. This is because with two copies of the UE^ s message, UE ₂ has the chance to decode the interference.

[0043] In the example, above Chase combining HARQ was used (i.e., Xi(t) = Xi (t)). In such cases, UE ₂ can use a maximum ratio combining receiver to decode X _± (t) from combining Y ₂ (t) and Y _z (t). However, the same setup is applicable for HARQ protocols other than Chase combining HARQ, such as, Incremental Redundancy HARQ, and UE ₂ need only use the appropriate receiver depending on the selected HARQ protocol. In summary, we can use the standard OMA-based receiver to decode X (t) and then, based on the SIC receiver concept, remove this signal to have an interference-free signal for UE ₂ .

[0044] Because UE ₂ may experience an interference-free condition in slot t, it is possible for the NN to adapt the transmission parameters for that time slot. Particularly, in slot t a higher data rate may be considered for UE _Z which leads to better fairness between UEi and UE ₂ . AS an alternative method, the NN may reduce the transmit power for the UE ₂ message which, with a sum power constraint R + P ₂ = P, increases the transmit power for UE- _L message and, consequently, increases its successful message decoding probability. That is, depending on the network designer’s point of interest, our proposed scheme can give the chance to reduce the error probability for UE- _L

[0045] The above process continues until UE ₁ correctly decodes its message or the maximum number of retransmission rounds is reached. In any round, if UE ₂ has correctly decoded the message of UE ₁ while UE ₁ has still not decoded its message, the interference signal can be easily removed by UE ₂ with no decoding process. Thus, UE ₂ can be provided with high data rates. Also, in any round, if UE ₂ can not correctly decode the message of UE _l it decodes-and-removes its own message and buffers the residual signal containing the message of UE ₁ for possible use in the next rounds. The key to enable such adaptive decoding scheme at UE ₂ is for UE ₂ to be aware that NN is retransmitting the message for UEi (e.g., UE ₂ being aware of the message decoding status of UE i). In some embodiments, the NN provides UE _Z with this information. In other embodiments, UE ₂ monitors the acknowledgements sent by UEi so that UE ₂ will know when UEi has sent a negative acknowledgment (NACK), and, thereby know that the NN will then retransmit the message to UEi.

[0046] To summarize, in one embodiment, the following steps may be performed:

[0047] Step 1 : The NN, using a NOMA transmission scheme, transmits in a first time slot (denoted slot tl here and denoted t above) a first signal S _ti , which contains a message for UEi (denoted Xi) and a message for UE ₂ (denoted X ₂ );

[0048] Step 2 : UEi receives signal Yi(tl) corresponding to Su (see equation (1)), UE ₂ receives signal Y ₂ (tl) corresponding to S _ti , UEi attempts to decode Xi from Yi(tl), and UE ₂ attempts to decode X ₂ from Y ₂ (tl). Assuming UEi was not able to decode Xi, UEi transmits a NACK (see block 302 in FIG. 3) so that NN will transmit in a following time slot (t2) (also referred to as t ) a second signal S _t2 that contains Xi and a message for UE ₂ (denoted X ₃ , which may be identical to X ₂ or a new message that is different than X ₂ ). Also, as described above, assuming UE ₂ was able to successfully decode X ₂ , UE ₂ generates the residual signal, as described above, and buffers the residual signal (see block 304 in FIG. 3). To synchronize the UEs’ signals, some delay may be considered by UE ₂ , shown by the“sleep” box 306 in FIG. 3 (alternatively the network may be the synchronization). [0049] Step 3: The NN receives the NACK transmitted by UEi. The NN also receives the acknowledgement sent by UE ₂ , which in this case we can assume is a positive

acknowledgement (ACK).

[0050] Step 4: The NN informs UE ₂ that in time slot t2 NN will transmit a signal (S _T2 ) comprising message X ₃ and message Xi (i.e., the message for UEi that UEi was not able to decode). The NN may also inform UEi of UE ₂ ’s decoding status (e.g., in this case NN informs UEi that UE ₂ was able to decode X ₂ ).

[0051] Step 5: in time slot t2 NN transmits S _t2 . Also, as described above, depending on the UEs message decoding status, the NN adapts the transmission parameters, e.g., rate and/or power, of the UEs. For instance, the power used to transmit message Xi in time slot t2 may be greater than the power that was used to transmit message Xi in time slot 1 and/or the rate used to transmit message X ₃ in time slot t2 may be greater than the rate used to transmit message X ₂ in time slot tl (i.e., message X ₃ contains more information than message X ₂ ).

[0052] Step 6: UE ₂ receives signal Y ₂ (t2) corresponding to S _t2 and UE ₂ first tries to decode Xi from signal Y ₂ (t2) using the buffered residual signal(s). If UE ₂ is successful in decoding Xi from signal Y ₂ (t2), then UE ₂ can remove Xi from signal Y ₂ (t2), thereby allowing UE ₂ to decode message X ₃ from signal Y ₂ (t2). That is, in each round, UE ₂ tries to decode the message of UE ₁ based on all interfering signals accumulated in different rounds. If UE ₂ is not successful to decode the message of UE^ it buffers the signals for possible uses in the next rounds. If it can decode the message of UE ₁ correctly, it removes the message of UE ₁ and decodes its own message interference -free.

[0053] The method above is described using Chase Combining HARQ, where the same signal is sent to UE ₁ in different retransmission rounds; however, as described above, the same setup is well applicable for, e.g., incremental redundancy HARQ, where different redundancy signals are sent in different retransmission rounds. For every type of HARQ, UE ₂ uses the standard decoding schemes to decode the message of UE ₁ based on the signals accumulated in different rounds. Further, the above description assumes that the channel coefficients remain constant in different time slots; however the same approach is applicable in the cases with different values of channel coefficients in successive rounds. [0054] B. The Adaptive Decoding Scheme for the strong UE (UEi)

[0055] In this embodiment, which is illustrated in FIG. 4, UEi utilizes the

acknowledge/negative acknowledgement (ACK/NACK) signals of UE ₂ to decide which receiver it will use. Consider two successive time slots, tl and t2. Using the above example where the NN transmits S _ti in time slot tl, there is reasonably high probability that the UE ₂ cannot correctly decode from S _ti message X ₂ because message Xi, which is also contained in S _ti appears as interference as well as because of the poor channel condition between UE ₂ and the NN. Thus, NN may need to retransmit X ₂ in timeslot t2.

[0056] UEi, on the other hand, can use SIC to correctly decode from S _ti both messages

Xi and X ₂ with high probability, as UEi experiences high channel quality. Then, in slot t2, when the NN retransmits the same message for the UE ₂ (i.e., X ₂ ) by transmitting signal S _t2 containing a message for UE1 (denoted X ₃ ) as well as X2, the UEi already has the decoded the message of the UE ₂ (i.e., X ₂ ). Thus, the UEi removes X ₂ from Yi(t2), with no decoding process, and UEi decodes its own interference-free message (X ₃ ) with conventional decoders.

[0057] In this way, the decoding complexity/delay and the error propagation probability is reduced, and the NOMA -based setup works in the same way as OMA -based systems. This process continues until the UE ₂ correctly decodes its message or the maximum number of retransmission rounds is reached. The key to enable such adaptive decoding scheme at the UEi is to know about the message decoding status, i.e., ACK/NACK feedback signals, of the UE ₂ . In some embodiment, the NN provides the UEi with this information or UEi obtains it itself. Finally, because the decoding delay of SIC and conventional receivers are different, both UEs synchronize their signals based on the decoding scheme of the UEi.

[0058] To summarize, in one embodiment, the following steps may be performed:

[0059] Step 1 : The NN, using a NOMA transmission scheme, transmits in time slot tl a first signal S _ti , which contains a message for UEi (denoted Xi) and a message for UE ₂ (denoted X ₂ );

[0060] Step 2 : UEi receives signal Y i(tl) corresponding to S _ti (see equation (1)), UE ₂ receives signal Y ₂ (tl) corresponding to S _ti , UEi attempts to decode Xi from Yi(tl), and UE ₂ attempts to decode X ₂ from Y ₂ (tl). Assuming UE ₂ was not able to decode X ₂ , UE ₂ transmits a NACK (see block 402 in FIG. 4), which, in this example, causes NN to transmit in a following time slot (t2) a second signal S _t 2 that contains X ₂ and a message for UEi (denoted X ₃ , which may be identical to Xi or a new message that is different than Xi).

[0061] Step 3: The NN receives the NACK transmitted by UE ₂ . The NN also receives the acknowledgement sent by UEi.

[0062] Step 4: The NN informs UEi of UE ₂ ’s decoding status, which informs UEi that in time slot t2 NN will transmit a signal (S _t2 ) comprising the message X ₃ for UEi and message X ₂ (i.e., the message for UE ₂ that UE ₂ was not able to decode). The NN may also inform UE ₂ of UEi’s decoding status. Assuming UEi transmits an ACK (positive acknowledgement), the NN will assume that in time slot t2 UEi will use a non-SIC decoder to decode message X ₃ (e.g., NN will assume that in time slot t2 both UEi and UE ₂ will use the same decoders or decoders that have the same timing) and the NN will inform UE ₂ such that the signaling procedure of the UEs are synchronized. This can be performed by, e.g., NN sending to UE ₂ information (e.g., an index) identifying a synchronization protocol (e.g., information indicating whether UE ₂ need to employ a sleeping period 401 such that after UE ₂ finishes its decoding process it waits for a certain amount of time before transmitting an acknowledgement).

[0063] Step 5: UEi, after receiving the information about the message decoding status of the UE ₂ , selects an appropriate decoding scheme, i.e., SIC or conventional receiver, in the next time slot and adapts its synchronization protocol accordingly. Also, UE ₂ , depending on the selected decoding scheme of the UEi, uses the received synchronization index to synchronize its transmit/receive timing.

[0064] Step 6: in time slot t2 NN transmits S _t2 . Also, as described above, depending on the UEs message decoding status, the NN adapts the transmission parameters, e.g., rate and/or power, of the UEs. For instance, the power used to transmit message X ₂ in time slot t2 may be greater than the power that was used to transmit message X ₂ in time slot 1 and/or the rate used to transmit message X ₃ in time slot t2 may be greater than the rate used to transmit message Xi in time slot tl (i.e., message X ₃ contains more information than message Xi).

[0065] Step 7: UEi receives signal Y i(t2) corresponding to S _t2 and UEi uses the selected decoding scheme to decode X ₃ from Yi(t2). For example, assuming that UEi was able to decode message X ₂ from Yi(tl), then when UEi receives Yi(t2) in time slot t2, UEi can remove X ₂ from Y i(t2) and then use a non-SIC receiver to decode message X3 from the residual (see block 404 in FIG. 4).

[0066] The effectiveness of the proposed scheme depends on the channel coefficients remaining generally constant during the retransmission rounds, so that knowing X ₂ , the received interference term H _1l [R _l ^~ C ₂ can be removed by UEi from the received signal Y i(t2). As a result, the proposed scheme is most useful in stationary/slow-moving networks which, due to the high CSI acquisition and scheduling overhead of NOMA, are of most interest in NOMA-based systems.

[0067] The method above is described using Chase Combining HARQ, where the same signal is sent to UE ₂ in different retransmission rounds; however, as described above, the same setup is well applicable for, e.g., incremental redundancy HARQ, where different redundancy signals are sent in different retransmission rounds. There, decoding the message of the UE ₂ , the UEi knows the redundancy signals sent for the UE ₂ in next retransmissions and, as a result, can remove their corresponding received signals. Also, while the methods described above are described for the simplest case where, while the UE cannot correctly decode its message in round 1 , the same procedure is applicable in any round of retransmission. The method described above use a scenario in which a strong UE is paired with a weak UE; this is motivated by the fact that the highest gain of NOMA is observed when two UEs with different channel qualities are paired. Nevertheless, the same setup is applicable for any pair of UEs.

[0068] FIG. 5 is a flow chart illustrating a process 500, according to an embodiment, that is performed by the network node. Process 500 may begin in step s502 where the network node transmits, during a first time slot (tl), a first superimposed signal (S(tl)) comprising a first message for a first (e.g., UE 101 or UE 102) UE and a second message for a second UE (e.g., UE 101 or UE 102). In step s504, the network node receives a message transmitted by the first UE indicating that the first UE was unable to decode the first message. In step s506, after receiving the message transmitted by the first UE indicating that the first UE was unable to decode the first message, the network node provides to the second UE a message indicating that the first UE was unable to decode the first message. [0069] In some embodiments, the process may further include, after receiving the message transmitted by the first UE indicating that the first UE was unable to decode the first message, the network node transmitting, during a second time slot (t2), a second superimposed signal (S(t2)) comprising the first message for the first UE and a third message for the second UE. Also, the process may further include the network node transmitting, during a third time slot, a third superimposed signal comprising a third message for the first UE and a fourth message for the second UE; the network node receiving a message transmitted by the second UE indicating that the second UE was unable to decode the fourth message for the second UE included in the third superimposed signal; and after receiving the message transmitted by the second UE indicating that the second UE was unable to decode the fourth message, the network node providing to the first UE a message indicating that the network node will retransmit in a fourth time slot at least a portion of the fourth message. The process may further include, after receiving the message transmitted by the second UE indicating that the second UE was unable to decode the fourth message, the network node transmitting, during the fourth time slot, a fourth superimposed signal comprising the fourth message for the second UE and a fifth message for the first UE.

[0070] FIG. 6 is a flow chart illustrating a process 600 according to an embodiment.

Process 600 may begin in step s602 where a first UE (e.g., UE 101 or UE 102) receives, during a first time slot, a first superimposed signal based on a superimposed signal transmitted by the network node, the first superimposed signal comprising a first message for the first UE and a second message for a second UE. In step s604, after receiving the first superimposed signal, the first UE obtains information that indicates that the second UE was unable to decode the second message and uses the obtained information to decode a third message for the first UE that is contained in a second superimposed signal received by the first UE. In some

embodiments, the UE obtains the information by receiving a message transmitted by the network node, wherein the message indicates that the second UE was unable to decode the second message. In another embodiment, the first UE obtains the information comprises receiving a negative acknowledgement transmitted by the second UE.

[0071] In some embodiments, the first UE is a weak UE (e.g., UE ₂ ) and further performs steps s606-s6l0. In step s606, the first UE: decodes the first message from the first superimposed signal, uses the obtained first message and the first superimposed signal to generate a residual signal comprising the second message for the second UE, and buffers the residual signal. In step s608 the first UE receives, during a second time slot, a second superimposed signal, the second superimposed signal comprising a third message for the first UE and at least a portion of the second message. In step s6l0, as a result of receiving the information indicating that the second UE was not able to decode the second message, the first UE: i) uses the buffered residual signal and the received second superimposed signal to obtain the second message (or portion thereof in case of incremental HARQ) and ii) uses the obtained second message (or portion thereof) and the received second superimposed signal to obtain the third message from the second superimposed signal.

[0072] In other embodiments, the first UE is a strong UE (e.g., UEi) and further performs steps s6l2-s6l6. In step s6l2, the first UE obtains the second message from the first superimposed signal. In step s6l4, the first UE receives, during a second time slot, a second superimposed signal, the second superimposed signal comprising a third message for the first UE and at least a portion of the second message. In step s6l6, as a result of receiving the information indicating that the second UE was not able to decode the second message, the first UE removes the at least a portion of the second message from the second superimposed signal (e.g., removes 1i _ΐL [R _z ^~ C ₂ from the second superimposed signal) to produce a residual signal ^ar| d then obtains the third message (X ₃ ) from the residual signal using, for example, a conventional non-SIC decoder.

[0073] FIG. 7 is a block diagram of a network node 700, according to some embodiments for performing methods disclosed herein. As shown in FIG. 7, network node 700 may comprise: processing circuitry (PC) 702, which may include one or more processors (P) 755 (e.g., a general purpose microprocessor and/or one or more other processors, such as an application specific integrated circuit (ASIC), field-programmable gate arrays (FPGAs), and the like); a network interface 748 comprising a transmitter (Tx) 745 and a receiver (Rx) 747 for enabling network node 700 to transmit data to and receive data from other nodes connected to a network 110 (e.g., an Internet Protocol (IP) network) to which network interface 748 is connected; circuitry 703 (e.g., radio transceiver circuitry comprising an Rx 705 and a Tx 706) coupled to an antenna system 704 for wireless communication with UEs); and a local storage unit (a.k.a.,“data storage system”) 708, which may include one or more non-volatile storage devices and/or one or more volatile storage devices. In embodiments where PC 702 includes a programmable processor, a computer program product (CPP) 741 may be provided. CPP 741 includes a computer readable medium (CRM) 742 storing a computer program (CP) 743 comprising computer readable instructions (CRI) 744. CRM 742 may be a non-transitory computer readable medium, such as, magnetic media (e.g., a hard disk), optical media, memory devices (e.g., random access memory, flash memory), and the like. In some embodiments, the CRI 744 of computer program 743 is configured such that when executed by PC 702, the CRI causes network node 700 to perform steps described herein (e.g., steps described herein with reference to the flow charts). In other embodiments, network node 700 may be configured to perform steps described herein without the need for code. That is, for example, PC 702 may consist merely of one or more ASICs.

Hence, the features of the embodiments described herein may be implemented in hardware and/or software.

[0074] FIG. 8 is a diagram showing functional units of network node 105 according to one embodiment. In the embodiment shown, network node 105 includes: a transmission unit 802 for employing a transmitter to transmit a first superimposed signal; a receiver unit 804 for employing a receiver to obtain a message transmitted by the first UE (e.g., UE 101 or UE 102) indicating that the first UE was unable to decode the first message; and a providing unit 806 for providing to the second UE a message indicating that the first UE was unable to decode the first message.

[0075] FIG. 9 is a block diagram of a UE (e.g. UE 101 or UE 102), according to some embodiments. As shown in FIG. 9, the UE may comprise: processing circuitry (PC) 902, which may include one or more processors (P) 955 (e.g., a general purpose microprocessor and/or one or more other processors, such as an application specific integrated circuit (ASIC), field- programmable gate arrays (FPGAs), and the like); circuitry 903 (e.g., radio transceiver circuitry comprising an Rx 905 and a Tx 906) coupled to an antenna system 904 for wireless

communication); and a local storage unit (a.k.a.,“data storage system”) 908, which may include one or more non-volatile storage devices and/or one or more volatile storage devices. In embodiments where PC 902 includes a programmable processor, a computer program product (CPP) 941 may be provided. CPP 941 includes a computer readable medium (CRM) 942 storing a computer program (CP) 943 comprising computer readable instructions (CRI) 944. CRM 942 may be a non-transitory computer readable medium, such as, magnetic media (e.g., a hard disk), optical media, memory devices (e.g., random access memory, flash memory), and the like. In some embodiments, the CRI 944 of computer program 943 is configured such that when executed by PC 902, the CRI causes the UE to perform steps described herein (e.g., steps described herein with reference to the flow charts). In other embodiments, the UE may be configured to perform steps described herein without the need for code. That is, for example, PC 902 may consist merely of one or more ASICs. Hence, the features of the embodiments described herein may be implemented in hardware and/or software.

[0076] FIG. 10 is a diagram showing functional units of a UE (e.g., UE 101 or UE 102) according to one embodiment. In the embodiment shown, the UE includes: a receiving unit 1002 for employing a receiver to obtain a first superimposed signal based on a superimposed signal transmitted by the network node, an information obtaining unit 1004 for obtaining information indicating that the second UE was unable to decode the second message from the first superimposed signal, and a decoding unit 1006 for using the obtained information to decode a third message for the first UE that is contained in a second superimposed signal received by the first UE.

[0077] FIG. 11 is a diagram showing functional units of UE 101 according to one embodiment. In the embodiment shown, UE 101 includes a first decoding unit 1102 for obtaining first and second messages from a first superimposed signal, wherein the first message is for the first UE and the second message is for a second UE (e.g., UE 102). UE 101 further includes an information obtaining unit 1104 for obtaining information that indicates that the second UE was unable to decode the second message from the first superimposed signal. UE 101 further includes a processing unit 1106 configured such that, as a result of the obtaining unit obtaining information that indicates that the second UE was unable to decode the second message from the first superimposed signal, the processing unit 1106 employs a second decoding unit 1108 to obtain a third message from a second superimposed signal comprising the third message and at least a portion of the second message. The second decoding unit 1108 is different than the first decoding unit 1102. The second decoding unit 1108 is configured to use the second message obtained via the first decoding unit 1102 to remove the at least a portion of the second message from the second superimposed signal, thereby producing a residual signal. And the second decoding unit 1108 is configured to obtain the third message from the residual signal.

[0078] While various embodiments are described herein, it should be understood that they have been presented by way of example only, and not limitation. Thus, the breadth and scope of this disclosure should not be limited by any of the above-described exemplary embodiments. Moreover, any combination of the above-described elements in all possible variations thereof is encompassed by the disclosure unless otherwise indicated herein or otherwise clearly contradicted by context.

[0079] Additionally, while the processes described above and illustrated in the drawings are shown as a sequence of steps, this was done solely for the sake of illustration. Accordingly, it is contemplated that some steps may be added, some steps may be omitted, the order of the steps may be re-arranged, and some steps may be performed in parallel.